Han Shao
Trustworthy Machine Learning under Social and Adversarial Data Sources
Research Abstract:
Machine learning has witnessed remarkable breakthroughs in recent years. Most machine learning techniques assume that the training and test data are sampled from an underlying distribution and aim to find a predictor with low population loss. However, data may be generated by strategic individuals, collected by self-interested agents, possibly poisoned by adversarial attackers, and used to create predictors, models, and policies satisfying multiple objectives. As a result, predictors may underperform. To ensure the success of machine learning, it is crucial to develop trustworthy algorithms capable of handling these factors. My research has been on modeling strategic and adversarial data sources, analyzing their effects on predictors, and designing new methods and insights to enhance accuracy.
Bio:
Han Shao is a fifth-year Ph.D. student at TTIC, advised by Prof. Avrim Blum. Her research focuses on machine learning theory. She has been working on modeling strategic behaviors of agents in the learning process and developing robust algorithms. She is also interested in gaining a theoretical understanding of empirical observations concerning robustness.